Intelligent Computational Framework for Adaptive Temperature and Humidity Control in Air Conditioning Systems

This study investigates the implementation of an intelligent AC control system using multiple methodologies to enhance energy efficiency and user comfort. Traditional control systems often rely on fixed algorithms, which can lead to suboptimal performance in varying environmental conditions. In this paper four distinct approaches were explored. A computational intelligence-based framework was designed using a Convolutional Neural Network (CNN) with the Adam optimizer which was validated using Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) integrated with Support Vector Machines (SVM) & Decision Tree (DT), and Fuzzy Logic which uses rule-based reasoning for control decisions. Each method was evaluated on metrics such as accuracy, mean absolute error (MAE), computational efficiency, and energy savings. The CNN with Adam optimizer achieved the highest accuracy and energy efficiency, while Fuzzy Logic provided fast and interpretable decisions. GA and PSO effectively optimized feature selection and parameters for machine learning models, offering balanced performance for resource-constrained systems.